March 21, 2016

SXSW 2016: AI, humanity and the future

By Tom Kelshaw Metalworks Director of Technology

Tom Kelshaw

Tom Kelshaw

Metalworks Director of Technology

South by Southwest Interactive 2016 presented a dichotomy of topics: hype for an AI-empowered future, and a long overdue look at humanity, diversity & fairness.

Part 1: Artificial Intelligence

Last year at SXSW 2015 we wrote about rise of the robots – AI, drones etc. This year, those robots have risen and are driving big conversations in Austin.

Talk of Artificial Intelligence – AI, Deep Learning, Automation, Robotics – was everywhere at SXSW. Kevin Kelly (Wired Magazine) highlighted AI as one of the top three technologies that would shape the next 30 years through “augmenting” humans. At the IBM Cognitive Studio, there were lines down the street to experience Watson AI demonstrations. And whilst Austin prepared for SXSW, over in Seoul Google’s AlphaGo AI system was smashing world champion Lee Se-dol in the game Go. Go was long-thought to be a milestone in AI – far more complex and creative than chess.

We attended sessions about AI from Stanford professors, MIT roboticists, German brand marketers, Dutch tech artists and workforce psychologists. Key themes included:


As AI evolves exponentially due to techniques such as Deep Learning and Neural Networks, 2016 will see more ‘€œvertical AIs’€ that can take over complex tasks in a single category – and maybe even entire jobs.

  • – creators of the popular ‘€œAmy’€ virtual assistant – performs the heinous task of calendar scheduling. Many people don’€™t even know Amy is a bot.
  • Digital Genius are creating specific AIs that can take over up to 60% of customer service chats.
  • Viv, from the creators of Siri, can already handle most voice calls for pre-sales or tech support.
  • Tara is an AI chatbot that interviews customers to determine their project needs and assign them to the correct freelance technical or design resources.

Tim Leberehct (frog design) spoke about the threat of displacing hundreds of millions of workers in the quest for productivity. He defines ‘€œThe Dispensibles’€ as those workers (up to 50% of us by 2025!) that might be replaced by sophisticated AIs. Similar to modern day refugees, they’€™ll have specific challenges and needs – how might marketers cope with a new audience with plenty of time on their hands, but limited means?


For Maxus clients, we’€™re already working with AI to drive creative outcomes. Rather than replacing creators, the use of algorithms and human-like understanding can augment our process. Our current Clarins USA campaign uses AI to understand images that best represent moments of joy & triumph.

Dutch artist Nick Verstand presented his amazing work ANIMA – an AI inside an 8ft glowing, swirling sphere. The planetary orb reacts to the audience near it, learning from their touches, motions, voices and reactions. It was mesmerizing. It’€™s hard to understand exactly what the AI was ‘€œthinking’€ and trying to achieve – but that’€™s the same with most art.

IBM Watson researchers spoke about their AI XPRIZE with TED conference. IBM is offering 5 million USD to the winning AI that can deliver the best TED talk in 2017, as chosen by the (human) audience.

Friends of Metalworks, Dave & Gabe Studio, collaborated with sound artists and coders to create an interactive installation at the Deloitte Digital ARC activation. Adapting to physical input from participants, the system used algorithmic methods to ensure an ongoing, organic and tonal synth track that sounded awesome.

Spanish startup TheMotion caught our eye with their algorithm-powered video advertisement creator. Point their tech at any e-commerce website, and it’€™ll gather all the required information to make video ads with pricing, offers. They can even be personalised based on Facebook or other data.


If robots perform more of our daily tasks, including creative ones, we’€™ll need to watch the amount of trust we place in their algorithms – and how those algorithms will affect brand trust.

Google spoke about the challenge of self-driving cars and the ‘€œtrolley problem‘€ – in an unavoidable collision, how might our car decide who to injure or even kill?

Instagram announced a shift to curated (rather than chronological) feed, meaning an algorithm will decide what you see. Instagram is foregoing the laws of nature (time) for those of the machine (algorithm). Will users be happy to trust an algorithm to determine which picture of someone else’s lunch they see?

Uber constantly comes under fire for its algorithmic ‘€œsurge pricing’€. Whilst it may work to get more cars where they’€™re needed, it can be seen as exploitative (e.g paying $400 on New Years Eve) or even callous (as in the case of the 2014 Sydney shooting, where people fleeing the area were subject to higher charges).

In an algorithmic world, making choices a little more sensitive will be important to keeping brands ‘€œhuman’€ and maintaining trust. More on the ‘€œhumanity’€ side of SXSW2016 in Part 2.